The shutdown while deployment/startup is occuring is a real mess at the moment this needs fixing
such that we get at most one error (there will always be race conditions between
internal service deployment and undeployment that are unexpected - but should be fixed as they are found)
and with better communication between the "layers" no errors.